AI as a modern Solon’s Compass: enhancing accuracy, efficiency and ethical judgment in complex commercial disputes

Abstract

Purpose This study explores how artificial intelligence (AI) assistance, likened to a modern “Solon’s Compass” – a metaphor inspired by Solon, the Athenian lawgiver known for wise and equitable reforms, symbolizing AI’sguidance toward precise, efficient and ethically aware decisions – impacts decision-making in complex commercial disputes, focusing on accuracy, efficiency and perceived ethical quality. It addresses gaps in understanding AI’srole in enhancing these dimensions through three pre-registered randomized experiments. Design/methodology/approach Three pre-registered randomized experiments were conducted. Study 1 involved university students resolving contract disputes with and without AI assistance. Study 2 used a diverse sample from Amazon Mechanical Turk to assess decision times and accuracy. Study 3 examined the cognitive mechanisms, including cognitive load, trust in AI and ethical decision-making. Findings The results reveal that AI assistance improves decision-making speed and increases perceived ethical quality, such as fairness and transparency, based on self-assessments. Participants with AI support reported higher perceptions of accuracy and perceived improvements in fairness, potentially influenced by their assumptions about AI’scapabilities, compared to those without assistance, highlighting AI’spotential to influence self-reported decision quality in ethically and legally complex scenarios. Originality/value This research contributes to the underexplored area of AI’srole in complex, ethically sensitive decision-making processes. By focusing on commercial disputes, it provides new insights into AI’spotential to improve efficiency and increases self-assessment about ethical aspects.

Key findings

• ▸AI assistance improves decision-making speed
• ▸AI assistance increases perceived ethical quality
• ▸AI assistance enhances accuracy in complex commercial disputes

Keywords

Identifiers

DOI

10.1108/k-12-2024-3243

Journal

Kybernetes

Year

2026